mirror of
https://github.com/ZDoom/gzdoom.git
synced 2024-11-22 04:01:31 +00:00
Lemon update 2014-01-11 03:06:18 on branch lemon-updates
- Add the new "%token_class" directive for defining symbolic names that stand any one of a collection of tokens. (user: drh)
This commit is contained in:
Randy Heit
parent
249aed455b
commit
e59ef08cc8
1 changed files with 71 additions and 19 deletions
|
|
@ -1463,12 +1463,15 @@ int main(int argc, char **argv)
|
|||
}
|
||||
|
||||
/* Count and index the symbols of the grammar */
|
||||
lem.nsymbol = Symbol_count();
|
||||
Symbol_new("{default}");
|
||||
lem.nsymbol = Symbol_count();
|
||||
lem.symbols = Symbol_arrayof();
|
||||
for(i=0; i<=lem.nsymbol; i++) lem.symbols[i]->index = i;
|
||||
qsort(lem.symbols,lem.nsymbol+1,sizeof(struct symbol*), Symbolcmpp);
|
||||
for(i=0; i<=lem.nsymbol; i++) lem.symbols[i]->index = i;
|
||||
for(i=0; i<lem.nsymbol; i++) lem.symbols[i]->index = i;
|
||||
qsort(lem.symbols,lem.nsymbol,sizeof(struct symbol*), Symbolcmpp);
|
||||
for(i=0; i<lem.nsymbol; i++) lem.symbols[i]->index = i;
|
||||
while( lem.symbols[i-1]->type==MULTITERMINAL ){ i--; }
|
||||
assert( strcmp(lem.symbols[i-1]->name,"{default}")==0 );
|
||||
lem.nsymbol = i - 1;
|
||||
for(i=1; isupper(lem.symbols[i]->name[0]); i++);
|
||||
lem.nterminal = i;
|
||||
|
||||
|
|
@ -1947,7 +1950,9 @@ enum e_state {
|
|||
WAITING_FOR_DESTRUCTOR_SYMBOL,
|
||||
WAITING_FOR_DATATYPE_SYMBOL,
|
||||
WAITING_FOR_FALLBACK_ID,
|
||||
WAITING_FOR_WILDCARD_ID
|
||||
WAITING_FOR_WILDCARD_ID,
|
||||
WAITING_FOR_CLASS_ID,
|
||||
WAITING_FOR_CLASS_TOKEN
|
||||
};
|
||||
struct pstate {
|
||||
char *filename; /* Name of the input file */
|
||||
|
|
@ -1957,6 +1962,7 @@ struct pstate {
|
|||
struct lemon *gp; /* Global state vector */
|
||||
enum e_state state; /* The state of the parser */
|
||||
struct symbol *fallback; /* The fallback token */
|
||||
struct symbol *tkclass; /* Token class symbol */
|
||||
struct symbol *lhs; /* Left-hand side of current rule */
|
||||
const char *lhsalias; /* Alias for the LHS */
|
||||
int nrhs; /* Number of right-hand side symbols seen */
|
||||
|
|
@ -2260,6 +2266,8 @@ to follow the previous rule.");
|
|||
psp->state = WAITING_FOR_FALLBACK_ID;
|
||||
}else if( strcmp(x,"wildcard")==0 ){
|
||||
psp->state = WAITING_FOR_WILDCARD_ID;
|
||||
}else if( strcmp(x,"token_class")==0 ){
|
||||
psp->state = WAITING_FOR_CLASS_ID;
|
||||
}else{
|
||||
ErrorMsg(psp->filename,psp->tokenlineno,
|
||||
"Unknown declaration keyword: \"%%%s\".",x);
|
||||
|
|
@ -2428,6 +2436,40 @@ to follow the previous rule.");
|
|||
}
|
||||
}
|
||||
break;
|
||||
case WAITING_FOR_CLASS_ID:
|
||||
if( !islower(x[0]) ){
|
||||
ErrorMsg(psp->filename, psp->tokenlineno,
|
||||
"%%token_class must be followed by an identifier: ", x);
|
||||
psp->errorcnt++;
|
||||
psp->state = RESYNC_AFTER_DECL_ERROR;
|
||||
}else if( Symbol_find(x) ){
|
||||
ErrorMsg(psp->filename, psp->tokenlineno,
|
||||
"Symbol \"%s\" already used", x);
|
||||
psp->errorcnt++;
|
||||
psp->state = RESYNC_AFTER_DECL_ERROR;
|
||||
}else{
|
||||
psp->tkclass = Symbol_new(x);
|
||||
psp->tkclass->type = MULTITERMINAL;
|
||||
psp->state = WAITING_FOR_CLASS_TOKEN;
|
||||
}
|
||||
break;
|
||||
case WAITING_FOR_CLASS_TOKEN:
|
||||
if( x[0]=='.' ){
|
||||
psp->state = WAITING_FOR_DECL_OR_RULE;
|
||||
}else if( isupper(x[0]) || ((x[0]=='|' || x[0]=='/') && isupper(x[1])) ){
|
||||
struct symbol *msp = psp->tkclass;
|
||||
msp->nsubsym++;
|
||||
msp->subsym = (struct symbol **) realloc(msp->subsym,
|
||||
sizeof(struct symbol*)*msp->nsubsym);
|
||||
if( !isupper(x[0]) ) x++;
|
||||
msp->subsym[msp->nsubsym-1] = Symbol_new(x);
|
||||
}else{
|
||||
ErrorMsg(psp->filename, psp->tokenlineno,
|
||||
"%%token_class argument \"%s\" should be a token", x);
|
||||
psp->errorcnt++;
|
||||
psp->state = RESYNC_AFTER_DECL_ERROR;
|
||||
}
|
||||
break;
|
||||
case RESYNC_AFTER_RULE_ERROR:
|
||||
/* if( x[0]=='.' ) psp->state = WAITING_FOR_DECL_OR_RULE;
|
||||
** break; */
|
||||
|
|
@ -2800,11 +2842,13 @@ void Reprint(struct lemon *lemp)
|
|||
printf(" ::=");
|
||||
for(i=0; i<rp->nrhs; i++){
|
||||
sp = rp->rhs[i];
|
||||
printf(" %s", sp->name);
|
||||
if( sp->type==MULTITERMINAL ){
|
||||
printf(" %s", sp->subsym[0]->name);
|
||||
for(j=1; j<sp->nsubsym; j++){
|
||||
printf("|%s", sp->subsym[j]->name);
|
||||
}
|
||||
}else{
|
||||
printf(" %s", sp->name);
|
||||
}
|
||||
/* if( rp->rhsalias[i] ) printf("(%s)",rp->rhsalias[i]); */
|
||||
}
|
||||
|
|
@ -2826,11 +2870,13 @@ void ConfigPrint(FILE *fp, struct config *cfp)
|
|||
if( i==cfp->dot ) fprintf(fp," *");
|
||||
if( i==rp->nrhs ) break;
|
||||
sp = rp->rhs[i];
|
||||
fprintf(fp," %s", sp->name);
|
||||
if( sp->type==MULTITERMINAL ){
|
||||
fprintf(fp," %s", sp->subsym[0]->name);
|
||||
for(j=1; j<sp->nsubsym; j++){
|
||||
fprintf(fp,"|%s",sp->subsym[j]->name);
|
||||
}
|
||||
}else{
|
||||
fprintf(fp," %s", sp->name);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -3587,9 +3633,11 @@ static void writeRuleText(FILE *out, struct rule *rp){
|
|||
fprintf(out,"%s ::=", rp->lhs->name);
|
||||
for(j=0; j<rp->nrhs; j++){
|
||||
struct symbol *sp = rp->rhs[j];
|
||||
if( sp->type!=MULTITERMINAL ){
|
||||
fprintf(out," %s", sp->name);
|
||||
if( sp->type==MULTITERMINAL ){
|
||||
}else{
|
||||
int k;
|
||||
fprintf(out," %s", sp->subsym[0]->name);
|
||||
for(k=1; k<sp->nsubsym; k++){
|
||||
fprintf(out,"|%s",sp->subsym[k]->name);
|
||||
}
|
||||
|
|
@ -4058,7 +4106,8 @@ void ReportHeader(struct lemon *lemp)
|
|||
if( in ){
|
||||
int nextChar;
|
||||
for(i=1; i<lemp->nterminal && fgets(line,LINESIZE,in); i++){
|
||||
sprintf(pattern,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
|
||||
sprintf(pattern,"#define %s%-30s %2d\n",
|
||||
prefix,lemp->symbols[i]->name,i);
|
||||
if( strcmp(line,pattern) ) break;
|
||||
}
|
||||
nextChar = fgetc(in);
|
||||
|
|
@ -4072,7 +4121,7 @@ void ReportHeader(struct lemon *lemp)
|
|||
out = file_open(lemp,".h","wb");
|
||||
if( out ){
|
||||
for(i=1; i<lemp->nterminal; i++){
|
||||
fprintf(out,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
|
||||
fprintf(out,"#define %s%-30s %3d\n",prefix,lemp->symbols[i]->name,i);
|
||||
}
|
||||
fclose(out);
|
||||
}
|
||||
|
|
@ -4442,11 +4491,15 @@ struct symbol *Symbol_new(const char *x)
|
|||
return sp;
|
||||
}
|
||||
|
||||
/* Compare two symbols for working purposes
|
||||
/* Compare two symbols for sorting purposes. Return negative,
|
||||
** zero, or positive if a is less then, equal to, or greater
|
||||
** than b.
|
||||
**
|
||||
** Symbols that begin with upper case letters (terminals or tokens)
|
||||
** must sort before symbols that begin with lower case letters
|
||||
** (non-terminals). Other than that, the order does not matter.
|
||||
** (non-terminals). And MULTITERMINAL symbols (created using the
|
||||
** %token_class directive) must sort at the very end. Other than
|
||||
** that, the order does not matter.
|
||||
**
|
||||
** We find experimentally that leaving the symbols in their original
|
||||
** order (the order they appeared in the grammar file) gives the
|
||||
|
|
@ -4454,12 +4507,11 @@ struct symbol *Symbol_new(const char *x)
|
|||
*/
|
||||
int Symbolcmpp(const void *_a, const void *_b)
|
||||
{
|
||||
const struct symbol **a = (const struct symbol **) _a;
|
||||
const struct symbol **b = (const struct symbol **) _b;
|
||||
int i1 = (**a).index + 10000000*((**a).name[0]>'Z');
|
||||
int i2 = (**b).index + 10000000*((**b).name[0]>'Z');
|
||||
assert( i1!=i2 || strcmp((**a).name,(**b).name)==0 );
|
||||
return i1-i2;
|
||||
const struct symbol *a = *(const struct symbol **) _a;
|
||||
const struct symbol *b = *(const struct symbol **) _b;
|
||||
int i1 = a->type==MULTITERMINAL ? 3 : a->name[0]>'Z' ? 2 : 1;
|
||||
int i2 = b->type==MULTITERMINAL ? 3 : b->name[0]>'Z' ? 2 : 1;
|
||||
return i1==i2 ? a->index - b->index : i1 - i2;
|
||||
}
|
||||
|
||||
/* There is one instance of the following structure for each
|
||||
|
|
|
|||
Loading…
Reference in a new issue